Auto upload 2026-08-07T22:11:20.527314
Browse files- out/mlp-linear-21L_run/checkpoint-700/config.json +35 -0
- out/mlp-linear-21L_run/checkpoint-700/model.safetensors +3 -0
- out/mlp-linear-21L_run/checkpoint-700/optimizer.pt +3 -0
- out/mlp-linear-21L_run/checkpoint-700/rng_state.pth +3 -0
- out/mlp-linear-21L_run/checkpoint-700/scheduler.pt +3 -0
- out/mlp-linear-21L_run/checkpoint-700/tokenizer.json +0 -0
- out/mlp-linear-21L_run/checkpoint-700/tokenizer_config.json +13 -0
- out/mlp-linear-21L_run/checkpoint-700/trainer_state.json +391 -0
- out/mlp-linear-21L_run/checkpoint-700/training_args.bin +3 -0
- out/mlp-linear-21L_run/checkpoint-800/config.json +35 -0
- out/mlp-linear-21L_run/checkpoint-800/model.safetensors +3 -0
- out/mlp-linear-21L_run/checkpoint-800/optimizer.pt +3 -0
- out/mlp-linear-21L_run/checkpoint-800/rng_state.pth +3 -0
- out/mlp-linear-21L_run/checkpoint-800/scheduler.pt +3 -0
- out/mlp-linear-21L_run/checkpoint-800/tokenizer.json +0 -0
- out/mlp-linear-21L_run/checkpoint-800/tokenizer_config.json +13 -0
- out/mlp-linear-21L_run/checkpoint-800/trainer_state.json +442 -0
- out/mlp-linear-21L_run/checkpoint-800/training_args.bin +3 -0
- out/mlp-linear-21L_run/checkpoint-900/config.json +35 -0
- out/mlp-linear-21L_run/checkpoint-900/model.safetensors +3 -0
- out/mlp-linear-21L_run/checkpoint-900/optimizer.pt +3 -0
- out/mlp-linear-21L_run/checkpoint-900/rng_state.pth +3 -0
- out/mlp-linear-21L_run/checkpoint-900/scheduler.pt +3 -0
- out/mlp-linear-21L_run/checkpoint-900/tokenizer.json +0 -0
- out/mlp-linear-21L_run/checkpoint-900/tokenizer_config.json +13 -0
- out/mlp-linear-21L_run/checkpoint-900/trainer_state.json +493 -0
- out/mlp-linear-21L_run/checkpoint-900/training_args.bin +3 -0
- out/mlp-linear-21L_run/model.safetensors +1 -1
- out/mlp-linear-21L_run/training_log.jsonl +24 -0
out/mlp-linear-21L_run/checkpoint-700/config.json
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation": "linear",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"TinyLlamaForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": false,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"bos_token_id": 1,
|
| 9 |
+
"dtype": "bfloat16",
|
| 10 |
+
"eos_token_id": 2,
|
| 11 |
+
"head_dim": 32,
|
| 12 |
+
"hidden_act": "silu",
|
| 13 |
+
"hidden_size": 128,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 256,
|
| 16 |
+
"max_position_embeddings": 512,
|
| 17 |
+
"mlp_bias": false,
|
| 18 |
+
"mlp_type": "mlp",
|
| 19 |
+
"model_type": "tiny_llama",
|
| 20 |
+
"num_attention_heads": 4,
|
| 21 |
+
"num_hidden_layers": 21,
|
| 22 |
+
"num_key_value_heads": 4,
|
| 23 |
+
"pad_token_id": 0,
|
| 24 |
+
"pretraining_tp": 1,
|
| 25 |
+
"rms_norm_eps": 1e-06,
|
| 26 |
+
"rope_parameters": {
|
| 27 |
+
"rope_theta": 10000.0,
|
| 28 |
+
"rope_type": "default"
|
| 29 |
+
},
|
| 30 |
+
"tie_word_embeddings": true,
|
| 31 |
+
"tokenizer_name": "w-ahmad/tiny-stories-tokenizer",
|
| 32 |
+
"transformers_version": "5.15.0.dev0",
|
| 33 |
+
"use_cache": false,
|
| 34 |
+
"vocab_size": 4096
|
| 35 |
+
}
|
out/mlp-linear-21L_run/checkpoint-700/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d967d9c31a1c3b2ffd67a1f530042be8fbc82254b56c3927dabce5e016105977
|
| 3 |
+
size 7959296
|
out/mlp-linear-21L_run/checkpoint-700/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c93ed69554527aeff58169a5ec39202bc5352eee4da7cd92045445432328a608
|
| 3 |
+
size 16023738
|
out/mlp-linear-21L_run/checkpoint-700/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c07c9483d2aaa0e0aa4859fa05bacc55a60e0f30ff9c95a3b76854e880483a96
|
| 3 |
+
size 14244
|
out/mlp-linear-21L_run/checkpoint-700/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3173cd6908322b02fdacbfeb79bbb6a1a1a40d218262cceb79ee8c11c4fa07d7
|
| 3 |
+
size 1064
|
out/mlp-linear-21L_run/checkpoint-700/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
out/mlp-linear-21L_run/checkpoint-700/tokenizer_config.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<|endoftext|>",
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"errors": "replace",
|
| 7 |
+
"is_local": false,
|
| 8 |
+
"local_files_only": false,
|
| 9 |
+
"model_max_length": 1024,
|
| 10 |
+
"pad_token": "<|endoftext|>",
|
| 11 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 12 |
+
"unk_token": "<|endoftext|>"
|
| 13 |
+
}
|
out/mlp-linear-21L_run/checkpoint-700/trainer_state.json
ADDED
|
@@ -0,0 +1,391 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.37735849056603776,
|
| 6 |
+
"eval_steps": 50,
|
| 7 |
+
"global_step": 700,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.01078167115902965,
|
| 14 |
+
"grad_norm": 1.0546875,
|
| 15 |
+
"learning_rate": 0.001,
|
| 16 |
+
"loss": 7.274208068847656,
|
| 17 |
+
"step": 20
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.0215633423180593,
|
| 21 |
+
"grad_norm": 1.53125,
|
| 22 |
+
"learning_rate": 0.001,
|
| 23 |
+
"loss": 5.9999839782714846,
|
| 24 |
+
"step": 40
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.026954177897574125,
|
| 28 |
+
"eval_loss": 5.570199966430664,
|
| 29 |
+
"eval_runtime": 8.819,
|
| 30 |
+
"eval_samples_per_second": 1080.281,
|
| 31 |
+
"eval_steps_per_second": 2.154,
|
| 32 |
+
"step": 50
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.03234501347708895,
|
| 36 |
+
"grad_norm": 0.88671875,
|
| 37 |
+
"learning_rate": 0.001,
|
| 38 |
+
"loss": 5.567345046997071,
|
| 39 |
+
"step": 60
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.0431266846361186,
|
| 43 |
+
"grad_norm": 1.0546875,
|
| 44 |
+
"learning_rate": 0.001,
|
| 45 |
+
"loss": 5.250301742553711,
|
| 46 |
+
"step": 80
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"epoch": 0.05390835579514825,
|
| 50 |
+
"grad_norm": 0.50390625,
|
| 51 |
+
"learning_rate": 0.001,
|
| 52 |
+
"loss": 4.9394378662109375,
|
| 53 |
+
"step": 100
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 0.05390835579514825,
|
| 57 |
+
"eval_loss": 4.784651279449463,
|
| 58 |
+
"eval_runtime": 8.8621,
|
| 59 |
+
"eval_samples_per_second": 1075.032,
|
| 60 |
+
"eval_steps_per_second": 2.144,
|
| 61 |
+
"step": 100
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"epoch": 0.0646900269541779,
|
| 65 |
+
"grad_norm": 0.58203125,
|
| 66 |
+
"learning_rate": 0.001,
|
| 67 |
+
"loss": 4.660997772216797,
|
| 68 |
+
"step": 120
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.07547169811320754,
|
| 72 |
+
"grad_norm": 1.5546875,
|
| 73 |
+
"learning_rate": 0.001,
|
| 74 |
+
"loss": 4.43031005859375,
|
| 75 |
+
"step": 140
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.08086253369272237,
|
| 79 |
+
"eval_loss": 4.252280235290527,
|
| 80 |
+
"eval_runtime": 8.9869,
|
| 81 |
+
"eval_samples_per_second": 1060.1,
|
| 82 |
+
"eval_steps_per_second": 2.114,
|
| 83 |
+
"step": 150
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 0.0862533692722372,
|
| 87 |
+
"grad_norm": 0.48828125,
|
| 88 |
+
"learning_rate": 0.001,
|
| 89 |
+
"loss": 4.2591907501220705,
|
| 90 |
+
"step": 160
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 0.09703504043126684,
|
| 94 |
+
"grad_norm": 0.91015625,
|
| 95 |
+
"learning_rate": 0.001,
|
| 96 |
+
"loss": 4.122463607788086,
|
| 97 |
+
"step": 180
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"epoch": 0.1078167115902965,
|
| 101 |
+
"grad_norm": 0.435546875,
|
| 102 |
+
"learning_rate": 0.001,
|
| 103 |
+
"loss": 3.9880271911621095,
|
| 104 |
+
"step": 200
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.1078167115902965,
|
| 108 |
+
"eval_loss": 3.9186038970947266,
|
| 109 |
+
"eval_runtime": 8.9327,
|
| 110 |
+
"eval_samples_per_second": 1066.53,
|
| 111 |
+
"eval_steps_per_second": 2.127,
|
| 112 |
+
"step": 200
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 0.11859838274932614,
|
| 116 |
+
"grad_norm": 0.73828125,
|
| 117 |
+
"learning_rate": 0.001,
|
| 118 |
+
"loss": 3.8731555938720703,
|
| 119 |
+
"step": 220
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 0.1293800539083558,
|
| 123 |
+
"grad_norm": 0.40234375,
|
| 124 |
+
"learning_rate": 0.001,
|
| 125 |
+
"loss": 3.786928176879883,
|
| 126 |
+
"step": 240
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 0.1347708894878706,
|
| 130 |
+
"eval_loss": 3.682241916656494,
|
| 131 |
+
"eval_runtime": 8.9206,
|
| 132 |
+
"eval_samples_per_second": 1067.975,
|
| 133 |
+
"eval_steps_per_second": 2.13,
|
| 134 |
+
"step": 250
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.14016172506738545,
|
| 138 |
+
"grad_norm": 0.84375,
|
| 139 |
+
"learning_rate": 0.001,
|
| 140 |
+
"loss": 3.6825389862060547,
|
| 141 |
+
"step": 260
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.1509433962264151,
|
| 145 |
+
"grad_norm": 0.51171875,
|
| 146 |
+
"learning_rate": 0.001,
|
| 147 |
+
"loss": 3.6361549377441404,
|
| 148 |
+
"step": 280
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 0.16172506738544473,
|
| 152 |
+
"grad_norm": 1.5703125,
|
| 153 |
+
"learning_rate": 0.001,
|
| 154 |
+
"loss": 3.551879119873047,
|
| 155 |
+
"step": 300
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 0.16172506738544473,
|
| 159 |
+
"eval_loss": 3.5237605571746826,
|
| 160 |
+
"eval_runtime": 9.0532,
|
| 161 |
+
"eval_samples_per_second": 1052.333,
|
| 162 |
+
"eval_steps_per_second": 2.099,
|
| 163 |
+
"step": 300
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.1725067385444744,
|
| 167 |
+
"grad_norm": 0.62109375,
|
| 168 |
+
"learning_rate": 0.001,
|
| 169 |
+
"loss": 3.526197814941406,
|
| 170 |
+
"step": 320
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.18328840970350405,
|
| 174 |
+
"grad_norm": 0.34765625,
|
| 175 |
+
"learning_rate": 0.001,
|
| 176 |
+
"loss": 3.44549446105957,
|
| 177 |
+
"step": 340
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.18867924528301888,
|
| 181 |
+
"eval_loss": 3.4025301933288574,
|
| 182 |
+
"eval_runtime": 9.1498,
|
| 183 |
+
"eval_samples_per_second": 1041.23,
|
| 184 |
+
"eval_steps_per_second": 2.077,
|
| 185 |
+
"step": 350
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.1940700808625337,
|
| 189 |
+
"grad_norm": 1.0234375,
|
| 190 |
+
"learning_rate": 0.001,
|
| 191 |
+
"loss": 3.4011425018310546,
|
| 192 |
+
"step": 360
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.20485175202156333,
|
| 196 |
+
"grad_norm": 0.6171875,
|
| 197 |
+
"learning_rate": 0.001,
|
| 198 |
+
"loss": 3.3427574157714846,
|
| 199 |
+
"step": 380
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.215633423180593,
|
| 203 |
+
"grad_norm": 0.5859375,
|
| 204 |
+
"learning_rate": 0.001,
|
| 205 |
+
"loss": 3.2958106994628906,
|
| 206 |
+
"step": 400
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.215633423180593,
|
| 210 |
+
"eval_loss": 3.2790729999542236,
|
| 211 |
+
"eval_runtime": 9.1929,
|
| 212 |
+
"eval_samples_per_second": 1036.342,
|
| 213 |
+
"eval_steps_per_second": 2.067,
|
| 214 |
+
"step": 400
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"epoch": 0.22641509433962265,
|
| 218 |
+
"grad_norm": 0.5234375,
|
| 219 |
+
"learning_rate": 0.001,
|
| 220 |
+
"loss": 3.2616317749023436,
|
| 221 |
+
"step": 420
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.2371967654986523,
|
| 225 |
+
"grad_norm": 0.5703125,
|
| 226 |
+
"learning_rate": 0.001,
|
| 227 |
+
"loss": 3.2149822235107424,
|
| 228 |
+
"step": 440
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 0.24258760107816713,
|
| 232 |
+
"eval_loss": 3.181051015853882,
|
| 233 |
+
"eval_runtime": 9.1212,
|
| 234 |
+
"eval_samples_per_second": 1044.485,
|
| 235 |
+
"eval_steps_per_second": 2.083,
|
| 236 |
+
"step": 450
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.24797843665768193,
|
| 240 |
+
"grad_norm": 1.0546875,
|
| 241 |
+
"learning_rate": 0.001,
|
| 242 |
+
"loss": 3.1823314666748046,
|
| 243 |
+
"step": 460
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.2587601078167116,
|
| 247 |
+
"grad_norm": 0.5390625,
|
| 248 |
+
"learning_rate": 0.001,
|
| 249 |
+
"loss": 3.1341114044189453,
|
| 250 |
+
"step": 480
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 0.2695417789757412,
|
| 254 |
+
"grad_norm": 0.9375,
|
| 255 |
+
"learning_rate": 0.001,
|
| 256 |
+
"loss": 3.101780128479004,
|
| 257 |
+
"step": 500
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 0.2695417789757412,
|
| 261 |
+
"eval_loss": 3.1055757999420166,
|
| 262 |
+
"eval_runtime": 9.2225,
|
| 263 |
+
"eval_samples_per_second": 1033.013,
|
| 264 |
+
"eval_steps_per_second": 2.06,
|
| 265 |
+
"step": 500
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.2803234501347709,
|
| 269 |
+
"grad_norm": 0.5078125,
|
| 270 |
+
"learning_rate": 0.001,
|
| 271 |
+
"loss": 3.0730880737304687,
|
| 272 |
+
"step": 520
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.29110512129380056,
|
| 276 |
+
"grad_norm": 0.609375,
|
| 277 |
+
"learning_rate": 0.001,
|
| 278 |
+
"loss": 3.0466753005981446,
|
| 279 |
+
"step": 540
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.29649595687331537,
|
| 283 |
+
"eval_loss": 3.022505283355713,
|
| 284 |
+
"eval_runtime": 9.1838,
|
| 285 |
+
"eval_samples_per_second": 1037.366,
|
| 286 |
+
"eval_steps_per_second": 2.069,
|
| 287 |
+
"step": 550
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 0.3018867924528302,
|
| 291 |
+
"grad_norm": 0.5,
|
| 292 |
+
"learning_rate": 0.001,
|
| 293 |
+
"loss": 3.0200361251831054,
|
| 294 |
+
"step": 560
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 0.31266846361185985,
|
| 298 |
+
"grad_norm": 0.73046875,
|
| 299 |
+
"learning_rate": 0.001,
|
| 300 |
+
"loss": 2.9901798248291014,
|
| 301 |
+
"step": 580
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 0.32345013477088946,
|
| 305 |
+
"grad_norm": 0.73046875,
|
| 306 |
+
"learning_rate": 0.001,
|
| 307 |
+
"loss": 2.958896255493164,
|
| 308 |
+
"step": 600
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.32345013477088946,
|
| 312 |
+
"eval_loss": 2.95377516746521,
|
| 313 |
+
"eval_runtime": 9.1618,
|
| 314 |
+
"eval_samples_per_second": 1039.857,
|
| 315 |
+
"eval_steps_per_second": 2.074,
|
| 316 |
+
"step": 600
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 0.33423180592991913,
|
| 320 |
+
"grad_norm": 0.78125,
|
| 321 |
+
"learning_rate": 0.001,
|
| 322 |
+
"loss": 2.952092933654785,
|
| 323 |
+
"step": 620
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 0.3450134770889488,
|
| 327 |
+
"grad_norm": 0.44921875,
|
| 328 |
+
"learning_rate": 0.001,
|
| 329 |
+
"loss": 2.9132663726806642,
|
| 330 |
+
"step": 640
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 0.3504043126684636,
|
| 334 |
+
"eval_loss": 2.905498504638672,
|
| 335 |
+
"eval_runtime": 9.1525,
|
| 336 |
+
"eval_samples_per_second": 1040.913,
|
| 337 |
+
"eval_steps_per_second": 2.076,
|
| 338 |
+
"step": 650
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.3557951482479784,
|
| 342 |
+
"grad_norm": 0.98828125,
|
| 343 |
+
"learning_rate": 0.001,
|
| 344 |
+
"loss": 2.900470161437988,
|
| 345 |
+
"step": 660
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.3665768194070081,
|
| 349 |
+
"grad_norm": 0.56640625,
|
| 350 |
+
"learning_rate": 0.001,
|
| 351 |
+
"loss": 2.8784067153930666,
|
| 352 |
+
"step": 680
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.37735849056603776,
|
| 356 |
+
"grad_norm": 0.92578125,
|
| 357 |
+
"learning_rate": 0.001,
|
| 358 |
+
"loss": 2.8517574310302733,
|
| 359 |
+
"step": 700
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.37735849056603776,
|
| 363 |
+
"eval_loss": 2.868436336517334,
|
| 364 |
+
"eval_runtime": 9.2031,
|
| 365 |
+
"eval_samples_per_second": 1035.193,
|
| 366 |
+
"eval_steps_per_second": 2.065,
|
| 367 |
+
"step": 700
|
| 368 |
+
}
|
| 369 |
+
],
|
| 370 |
+
"logging_steps": 20,
|
| 371 |
+
"max_steps": 1000,
|
| 372 |
+
"num_input_tokens_seen": 0,
|
| 373 |
+
"num_train_epochs": 1,
|
| 374 |
+
"save_steps": 100,
|
| 375 |
+
"stateful_callbacks": {
|
| 376 |
+
"TrainerControl": {
|
| 377 |
+
"args": {
|
| 378 |
+
"should_epoch_stop": false,
|
| 379 |
+
"should_evaluate": false,
|
| 380 |
+
"should_log": false,
|
| 381 |
+
"should_save": true,
|
| 382 |
+
"should_training_stop": false
|
| 383 |
+
},
|
| 384 |
+
"attributes": {}
|
| 385 |
+
}
|
| 386 |
+
},
|
| 387 |
+
"total_flos": 3794221085491200.0,
|
| 388 |
+
"train_batch_size": 512,
|
| 389 |
+
"trial_name": null,
|
| 390 |
+
"trial_params": null
|
| 391 |
+
}
|
out/mlp-linear-21L_run/checkpoint-700/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f710638f98cb9aa723847c3078af1df019ed60911a4c32cfe8698e762841d517
|
| 3 |
+
size 4920
|
out/mlp-linear-21L_run/checkpoint-800/config.json
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation": "linear",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"TinyLlamaForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": false,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"bos_token_id": 1,
|
| 9 |
+
"dtype": "bfloat16",
|
| 10 |
+
"eos_token_id": 2,
|
| 11 |
+
"head_dim": 32,
|
| 12 |
+
"hidden_act": "silu",
|
| 13 |
+
"hidden_size": 128,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 256,
|
| 16 |
+
"max_position_embeddings": 512,
|
| 17 |
+
"mlp_bias": false,
|
| 18 |
+
"mlp_type": "mlp",
|
| 19 |
+
"model_type": "tiny_llama",
|
| 20 |
+
"num_attention_heads": 4,
|
| 21 |
+
"num_hidden_layers": 21,
|
| 22 |
+
"num_key_value_heads": 4,
|
| 23 |
+
"pad_token_id": 0,
|
| 24 |
+
"pretraining_tp": 1,
|
| 25 |
+
"rms_norm_eps": 1e-06,
|
| 26 |
+
"rope_parameters": {
|
| 27 |
+
"rope_theta": 10000.0,
|
| 28 |
+
"rope_type": "default"
|
| 29 |
+
},
|
| 30 |
+
"tie_word_embeddings": true,
|
| 31 |
+
"tokenizer_name": "w-ahmad/tiny-stories-tokenizer",
|
| 32 |
+
"transformers_version": "5.15.0.dev0",
|
| 33 |
+
"use_cache": false,
|
| 34 |
+
"vocab_size": 4096
|
| 35 |
+
}
|
out/mlp-linear-21L_run/checkpoint-800/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ad94815883a041eb8e3936f791ca66615dda183c4d79ebc751778bc6e063a2e2
|
| 3 |
+
size 7959296
|
out/mlp-linear-21L_run/checkpoint-800/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:62b7c82e614d5572ffa3b2681bcfe45b8a3c61cde3aa0de535f6c20e1537a692
|
| 3 |
+
size 16023738
|
out/mlp-linear-21L_run/checkpoint-800/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f83fe17529e572dff2734bb21512b28dd7cf5d20ef0e84688f5068ffbf24e765
|
| 3 |
+
size 14244
|
out/mlp-linear-21L_run/checkpoint-800/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7334026bbd250befa466dcbe6bec8bfe84c03024263e4dea55e40698f2caa92e
|
| 3 |
+
size 1064
|
out/mlp-linear-21L_run/checkpoint-800/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
out/mlp-linear-21L_run/checkpoint-800/tokenizer_config.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<|endoftext|>",
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"errors": "replace",
|
| 7 |
+
"is_local": false,
|
| 8 |
+
"local_files_only": false,
|
| 9 |
+
"model_max_length": 1024,
|
| 10 |
+
"pad_token": "<|endoftext|>",
|
| 11 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 12 |
+
"unk_token": "<|endoftext|>"
|
| 13 |
+
}
|
out/mlp-linear-21L_run/checkpoint-800/trainer_state.json
ADDED
|
@@ -0,0 +1,442 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.431266846361186,
|
| 6 |
+
"eval_steps": 50,
|
| 7 |
+
"global_step": 800,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.01078167115902965,
|
| 14 |
+
"grad_norm": 1.0546875,
|
| 15 |
+
"learning_rate": 0.001,
|
| 16 |
+
"loss": 7.274208068847656,
|
| 17 |
+
"step": 20
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.0215633423180593,
|
| 21 |
+
"grad_norm": 1.53125,
|
| 22 |
+
"learning_rate": 0.001,
|
| 23 |
+
"loss": 5.9999839782714846,
|
| 24 |
+
"step": 40
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.026954177897574125,
|
| 28 |
+
"eval_loss": 5.570199966430664,
|
| 29 |
+
"eval_runtime": 8.819,
|
| 30 |
+
"eval_samples_per_second": 1080.281,
|
| 31 |
+
"eval_steps_per_second": 2.154,
|
| 32 |
+
"step": 50
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.03234501347708895,
|
| 36 |
+
"grad_norm": 0.88671875,
|
| 37 |
+
"learning_rate": 0.001,
|
| 38 |
+
"loss": 5.567345046997071,
|
| 39 |
+
"step": 60
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.0431266846361186,
|
| 43 |
+
"grad_norm": 1.0546875,
|
| 44 |
+
"learning_rate": 0.001,
|
| 45 |
+
"loss": 5.250301742553711,
|
| 46 |
+
"step": 80
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"epoch": 0.05390835579514825,
|
| 50 |
+
"grad_norm": 0.50390625,
|
| 51 |
+
"learning_rate": 0.001,
|
| 52 |
+
"loss": 4.9394378662109375,
|
| 53 |
+
"step": 100
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 0.05390835579514825,
|
| 57 |
+
"eval_loss": 4.784651279449463,
|
| 58 |
+
"eval_runtime": 8.8621,
|
| 59 |
+
"eval_samples_per_second": 1075.032,
|
| 60 |
+
"eval_steps_per_second": 2.144,
|
| 61 |
+
"step": 100
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"epoch": 0.0646900269541779,
|
| 65 |
+
"grad_norm": 0.58203125,
|
| 66 |
+
"learning_rate": 0.001,
|
| 67 |
+
"loss": 4.660997772216797,
|
| 68 |
+
"step": 120
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.07547169811320754,
|
| 72 |
+
"grad_norm": 1.5546875,
|
| 73 |
+
"learning_rate": 0.001,
|
| 74 |
+
"loss": 4.43031005859375,
|
| 75 |
+
"step": 140
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.08086253369272237,
|
| 79 |
+
"eval_loss": 4.252280235290527,
|
| 80 |
+
"eval_runtime": 8.9869,
|
| 81 |
+
"eval_samples_per_second": 1060.1,
|
| 82 |
+
"eval_steps_per_second": 2.114,
|
| 83 |
+
"step": 150
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 0.0862533692722372,
|
| 87 |
+
"grad_norm": 0.48828125,
|
| 88 |
+
"learning_rate": 0.001,
|
| 89 |
+
"loss": 4.2591907501220705,
|
| 90 |
+
"step": 160
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 0.09703504043126684,
|
| 94 |
+
"grad_norm": 0.91015625,
|
| 95 |
+
"learning_rate": 0.001,
|
| 96 |
+
"loss": 4.122463607788086,
|
| 97 |
+
"step": 180
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"epoch": 0.1078167115902965,
|
| 101 |
+
"grad_norm": 0.435546875,
|
| 102 |
+
"learning_rate": 0.001,
|
| 103 |
+
"loss": 3.9880271911621095,
|
| 104 |
+
"step": 200
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.1078167115902965,
|
| 108 |
+
"eval_loss": 3.9186038970947266,
|
| 109 |
+
"eval_runtime": 8.9327,
|
| 110 |
+
"eval_samples_per_second": 1066.53,
|
| 111 |
+
"eval_steps_per_second": 2.127,
|
| 112 |
+
"step": 200
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 0.11859838274932614,
|
| 116 |
+
"grad_norm": 0.73828125,
|
| 117 |
+
"learning_rate": 0.001,
|
| 118 |
+
"loss": 3.8731555938720703,
|
| 119 |
+
"step": 220
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 0.1293800539083558,
|
| 123 |
+
"grad_norm": 0.40234375,
|
| 124 |
+
"learning_rate": 0.001,
|
| 125 |
+
"loss": 3.786928176879883,
|
| 126 |
+
"step": 240
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 0.1347708894878706,
|
| 130 |
+
"eval_loss": 3.682241916656494,
|
| 131 |
+
"eval_runtime": 8.9206,
|
| 132 |
+
"eval_samples_per_second": 1067.975,
|
| 133 |
+
"eval_steps_per_second": 2.13,
|
| 134 |
+
"step": 250
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.14016172506738545,
|
| 138 |
+
"grad_norm": 0.84375,
|
| 139 |
+
"learning_rate": 0.001,
|
| 140 |
+
"loss": 3.6825389862060547,
|
| 141 |
+
"step": 260
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.1509433962264151,
|
| 145 |
+
"grad_norm": 0.51171875,
|
| 146 |
+
"learning_rate": 0.001,
|
| 147 |
+
"loss": 3.6361549377441404,
|
| 148 |
+
"step": 280
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 0.16172506738544473,
|
| 152 |
+
"grad_norm": 1.5703125,
|
| 153 |
+
"learning_rate": 0.001,
|
| 154 |
+
"loss": 3.551879119873047,
|
| 155 |
+
"step": 300
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 0.16172506738544473,
|
| 159 |
+
"eval_loss": 3.5237605571746826,
|
| 160 |
+
"eval_runtime": 9.0532,
|
| 161 |
+
"eval_samples_per_second": 1052.333,
|
| 162 |
+
"eval_steps_per_second": 2.099,
|
| 163 |
+
"step": 300
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.1725067385444744,
|
| 167 |
+
"grad_norm": 0.62109375,
|
| 168 |
+
"learning_rate": 0.001,
|
| 169 |
+
"loss": 3.526197814941406,
|
| 170 |
+
"step": 320
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.18328840970350405,
|
| 174 |
+
"grad_norm": 0.34765625,
|
| 175 |
+
"learning_rate": 0.001,
|
| 176 |
+
"loss": 3.44549446105957,
|
| 177 |
+
"step": 340
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.18867924528301888,
|
| 181 |
+
"eval_loss": 3.4025301933288574,
|
| 182 |
+
"eval_runtime": 9.1498,
|
| 183 |
+
"eval_samples_per_second": 1041.23,
|
| 184 |
+
"eval_steps_per_second": 2.077,
|
| 185 |
+
"step": 350
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.1940700808625337,
|
| 189 |
+
"grad_norm": 1.0234375,
|
| 190 |
+
"learning_rate": 0.001,
|
| 191 |
+
"loss": 3.4011425018310546,
|
| 192 |
+
"step": 360
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.20485175202156333,
|
| 196 |
+
"grad_norm": 0.6171875,
|
| 197 |
+
"learning_rate": 0.001,
|
| 198 |
+
"loss": 3.3427574157714846,
|
| 199 |
+
"step": 380
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.215633423180593,
|
| 203 |
+
"grad_norm": 0.5859375,
|
| 204 |
+
"learning_rate": 0.001,
|
| 205 |
+
"loss": 3.2958106994628906,
|
| 206 |
+
"step": 400
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.215633423180593,
|
| 210 |
+
"eval_loss": 3.2790729999542236,
|
| 211 |
+
"eval_runtime": 9.1929,
|
| 212 |
+
"eval_samples_per_second": 1036.342,
|
| 213 |
+
"eval_steps_per_second": 2.067,
|
| 214 |
+
"step": 400
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"epoch": 0.22641509433962265,
|
| 218 |
+
"grad_norm": 0.5234375,
|
| 219 |
+
"learning_rate": 0.001,
|
| 220 |
+
"loss": 3.2616317749023436,
|
| 221 |
+
"step": 420
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.2371967654986523,
|
| 225 |
+
"grad_norm": 0.5703125,
|
| 226 |
+
"learning_rate": 0.001,
|
| 227 |
+
"loss": 3.2149822235107424,
|
| 228 |
+
"step": 440
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 0.24258760107816713,
|
| 232 |
+
"eval_loss": 3.181051015853882,
|
| 233 |
+
"eval_runtime": 9.1212,
|
| 234 |
+
"eval_samples_per_second": 1044.485,
|
| 235 |
+
"eval_steps_per_second": 2.083,
|
| 236 |
+
"step": 450
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.24797843665768193,
|
| 240 |
+
"grad_norm": 1.0546875,
|
| 241 |
+
"learning_rate": 0.001,
|
| 242 |
+
"loss": 3.1823314666748046,
|
| 243 |
+
"step": 460
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.2587601078167116,
|
| 247 |
+
"grad_norm": 0.5390625,
|
| 248 |
+
"learning_rate": 0.001,
|
| 249 |
+
"loss": 3.1341114044189453,
|
| 250 |
+
"step": 480
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 0.2695417789757412,
|
| 254 |
+
"grad_norm": 0.9375,
|
| 255 |
+
"learning_rate": 0.001,
|
| 256 |
+
"loss": 3.101780128479004,
|
| 257 |
+
"step": 500
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 0.2695417789757412,
|
| 261 |
+
"eval_loss": 3.1055757999420166,
|
| 262 |
+
"eval_runtime": 9.2225,
|
| 263 |
+
"eval_samples_per_second": 1033.013,
|
| 264 |
+
"eval_steps_per_second": 2.06,
|
| 265 |
+
"step": 500
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.2803234501347709,
|
| 269 |
+
"grad_norm": 0.5078125,
|
| 270 |
+
"learning_rate": 0.001,
|
| 271 |
+
"loss": 3.0730880737304687,
|
| 272 |
+
"step": 520
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.29110512129380056,
|
| 276 |
+
"grad_norm": 0.609375,
|
| 277 |
+
"learning_rate": 0.001,
|
| 278 |
+
"loss": 3.0466753005981446,
|
| 279 |
+
"step": 540
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.29649595687331537,
|
| 283 |
+
"eval_loss": 3.022505283355713,
|
| 284 |
+
"eval_runtime": 9.1838,
|
| 285 |
+
"eval_samples_per_second": 1037.366,
|
| 286 |
+
"eval_steps_per_second": 2.069,
|
| 287 |
+
"step": 550
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 0.3018867924528302,
|
| 291 |
+
"grad_norm": 0.5,
|
| 292 |
+
"learning_rate": 0.001,
|
| 293 |
+
"loss": 3.0200361251831054,
|
| 294 |
+
"step": 560
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 0.31266846361185985,
|
| 298 |
+
"grad_norm": 0.73046875,
|
| 299 |
+
"learning_rate": 0.001,
|
| 300 |
+
"loss": 2.9901798248291014,
|
| 301 |
+
"step": 580
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 0.32345013477088946,
|
| 305 |
+
"grad_norm": 0.73046875,
|
| 306 |
+
"learning_rate": 0.001,
|
| 307 |
+
"loss": 2.958896255493164,
|
| 308 |
+
"step": 600
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.32345013477088946,
|
| 312 |
+
"eval_loss": 2.95377516746521,
|
| 313 |
+
"eval_runtime": 9.1618,
|
| 314 |
+
"eval_samples_per_second": 1039.857,
|
| 315 |
+
"eval_steps_per_second": 2.074,
|
| 316 |
+
"step": 600
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 0.33423180592991913,
|
| 320 |
+
"grad_norm": 0.78125,
|
| 321 |
+
"learning_rate": 0.001,
|
| 322 |
+
"loss": 2.952092933654785,
|
| 323 |
+
"step": 620
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 0.3450134770889488,
|
| 327 |
+
"grad_norm": 0.44921875,
|
| 328 |
+
"learning_rate": 0.001,
|
| 329 |
+
"loss": 2.9132663726806642,
|
| 330 |
+
"step": 640
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 0.3504043126684636,
|
| 334 |
+
"eval_loss": 2.905498504638672,
|
| 335 |
+
"eval_runtime": 9.1525,
|
| 336 |
+
"eval_samples_per_second": 1040.913,
|
| 337 |
+
"eval_steps_per_second": 2.076,
|
| 338 |
+
"step": 650
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.3557951482479784,
|
| 342 |
+
"grad_norm": 0.98828125,
|
| 343 |
+
"learning_rate": 0.001,
|
| 344 |
+
"loss": 2.900470161437988,
|
| 345 |
+
"step": 660
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.3665768194070081,
|
| 349 |
+
"grad_norm": 0.56640625,
|
| 350 |
+
"learning_rate": 0.001,
|
| 351 |
+
"loss": 2.8784067153930666,
|
| 352 |
+
"step": 680
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.37735849056603776,
|
| 356 |
+
"grad_norm": 0.92578125,
|
| 357 |
+
"learning_rate": 0.001,
|
| 358 |
+
"loss": 2.8517574310302733,
|
| 359 |
+
"step": 700
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.37735849056603776,
|
| 363 |
+
"eval_loss": 2.868436336517334,
|
| 364 |
+
"eval_runtime": 9.2031,
|
| 365 |
+
"eval_samples_per_second": 1035.193,
|
| 366 |
+
"eval_steps_per_second": 2.065,
|
| 367 |
+
"step": 700
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.3881401617250674,
|
| 371 |
+
"grad_norm": 0.486328125,
|
| 372 |
+
"learning_rate": 0.001,
|
| 373 |
+
"loss": 2.846698760986328,
|
| 374 |
+
"step": 720
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.39892183288409705,
|
| 378 |
+
"grad_norm": 0.82421875,
|
| 379 |
+
"learning_rate": 0.001,
|
| 380 |
+
"loss": 2.8154937744140627,
|
| 381 |
+
"step": 740
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.40431266846361186,
|
| 385 |
+
"eval_loss": 2.813324451446533,
|
| 386 |
+
"eval_runtime": 9.1186,
|
| 387 |
+
"eval_samples_per_second": 1044.786,
|
| 388 |
+
"eval_steps_per_second": 2.084,
|
| 389 |
+
"step": 750
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 0.40970350404312667,
|
| 393 |
+
"grad_norm": 0.54296875,
|
| 394 |
+
"learning_rate": 0.001,
|
| 395 |
+
"loss": 2.8071964263916014,
|
| 396 |
+
"step": 760
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 0.42048517520215634,
|
| 400 |
+
"grad_norm": 0.69140625,
|
| 401 |
+
"learning_rate": 0.001,
|
| 402 |
+
"loss": 2.7811683654785155,
|
| 403 |
+
"step": 780
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 0.431266846361186,
|
| 407 |
+
"grad_norm": 0.73828125,
|
| 408 |
+
"learning_rate": 0.001,
|
| 409 |
+
"loss": 2.782184028625488,
|
| 410 |
+
"step": 800
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.431266846361186,
|
| 414 |
+
"eval_loss": 2.778449535369873,
|
| 415 |
+
"eval_runtime": 9.2351,
|
| 416 |
+
"eval_samples_per_second": 1031.606,
|
| 417 |
+
"eval_steps_per_second": 2.057,
|
| 418 |
+
"step": 800
|
| 419 |
+
}
|
| 420 |
+
],
|
| 421 |
+
"logging_steps": 20,
|
| 422 |
+
"max_steps": 1000,
|
| 423 |
+
"num_input_tokens_seen": 0,
|
| 424 |
+
"num_train_epochs": 1,
|
| 425 |
+
"save_steps": 100,
|
| 426 |
+
"stateful_callbacks": {
|
| 427 |
+
"TrainerControl": {
|
| 428 |
+
"args": {
|
| 429 |
+
"should_epoch_stop": false,
|
| 430 |
+
"should_evaluate": false,
|
| 431 |
+
"should_log": false,
|
| 432 |
+
"should_save": true,
|
| 433 |
+
"should_training_stop": false
|
| 434 |
+
},
|
| 435 |
+
"attributes": {}
|
| 436 |
+
}
|
| 437 |
+
},
|
| 438 |
+
"total_flos": 4336252669132800.0,
|
| 439 |
+
"train_batch_size": 512,
|
| 440 |
+
"trial_name": null,
|
| 441 |
+
"trial_params": null
|
| 442 |
+
}
|
out/mlp-linear-21L_run/checkpoint-800/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f710638f98cb9aa723847c3078af1df019ed60911a4c32cfe8698e762841d517
|
| 3 |
+
size 4920
|
out/mlp-linear-21L_run/checkpoint-900/config.json
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation": "linear",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"TinyLlamaForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": false,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"bos_token_id": 1,
|
| 9 |
+
"dtype": "bfloat16",
|
| 10 |
+
"eos_token_id": 2,
|
| 11 |
+
"head_dim": 32,
|
| 12 |
+
"hidden_act": "silu",
|
| 13 |
+
"hidden_size": 128,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 256,
|
| 16 |
+
"max_position_embeddings": 512,
|
| 17 |
+
"mlp_bias": false,
|
| 18 |
+
"mlp_type": "mlp",
|
| 19 |
+
"model_type": "tiny_llama",
|
| 20 |
+
"num_attention_heads": 4,
|
| 21 |
+
"num_hidden_layers": 21,
|
| 22 |
+
"num_key_value_heads": 4,
|
| 23 |
+
"pad_token_id": 0,
|
| 24 |
+
"pretraining_tp": 1,
|
| 25 |
+
"rms_norm_eps": 1e-06,
|
| 26 |
+
"rope_parameters": {
|
| 27 |
+
"rope_theta": 10000.0,
|
| 28 |
+
"rope_type": "default"
|
| 29 |
+
},
|
| 30 |
+
"tie_word_embeddings": true,
|
| 31 |
+
"tokenizer_name": "w-ahmad/tiny-stories-tokenizer",
|
| 32 |
+
"transformers_version": "5.15.0.dev0",
|
| 33 |
+
"use_cache": false,
|
| 34 |
+
"vocab_size": 4096
|
| 35 |
+
}
|
out/mlp-linear-21L_run/checkpoint-900/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:23ae85e4a038e289ca876dca6ccf4fd8a2a5cc4c9d17dd3a1d0dfe677c6fb506
|
| 3 |
+
size 7959296
|
out/mlp-linear-21L_run/checkpoint-900/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bd6450aae04a04bfe48b421dfd178026112bcd8d5af4497eb791228347c88c01
|
| 3 |
+
size 16023738
|
out/mlp-linear-21L_run/checkpoint-900/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:86a2aa7bca84d2b8e7dd4e04a286714ba9169af11c46c739950a52df4c45259f
|
| 3 |
+
size 14244
|
out/mlp-linear-21L_run/checkpoint-900/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6d48b00fea613f19335b87fb622648d28d7f8900963d07ba43c75df2b4dc3d4e
|
| 3 |
+
size 1064
|
out/mlp-linear-21L_run/checkpoint-900/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
out/mlp-linear-21L_run/checkpoint-900/tokenizer_config.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<|endoftext|>",
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"errors": "replace",
|
| 7 |
+
"is_local": false,
|
| 8 |
+
"local_files_only": false,
|
| 9 |
+
"model_max_length": 1024,
|
| 10 |
+
"pad_token": "<|endoftext|>",
|
| 11 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 12 |
+
"unk_token": "<|endoftext|>"
|
| 13 |
+
}
|
out/mlp-linear-21L_run/checkpoint-900/trainer_state.json
ADDED
|
@@ -0,0 +1,493 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.48517520215633425,
|
| 6 |
+
"eval_steps": 50,
|
| 7 |
+
"global_step": 900,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.01078167115902965,
|
| 14 |
+
"grad_norm": 1.0546875,
|
| 15 |
+
"learning_rate": 0.001,
|
| 16 |
+
"loss": 7.274208068847656,
|
| 17 |
+
"step": 20
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.0215633423180593,
|
| 21 |
+
"grad_norm": 1.53125,
|
| 22 |
+
"learning_rate": 0.001,
|
| 23 |
+
"loss": 5.9999839782714846,
|
| 24 |
+
"step": 40
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.026954177897574125,
|
| 28 |
+
"eval_loss": 5.570199966430664,
|
| 29 |
+
"eval_runtime": 8.819,
|
| 30 |
+
"eval_samples_per_second": 1080.281,
|
| 31 |
+
"eval_steps_per_second": 2.154,
|
| 32 |
+
"step": 50
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.03234501347708895,
|
| 36 |
+
"grad_norm": 0.88671875,
|
| 37 |
+
"learning_rate": 0.001,
|
| 38 |
+
"loss": 5.567345046997071,
|
| 39 |
+
"step": 60
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.0431266846361186,
|
| 43 |
+
"grad_norm": 1.0546875,
|
| 44 |
+
"learning_rate": 0.001,
|
| 45 |
+
"loss": 5.250301742553711,
|
| 46 |
+
"step": 80
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"epoch": 0.05390835579514825,
|
| 50 |
+
"grad_norm": 0.50390625,
|
| 51 |
+
"learning_rate": 0.001,
|
| 52 |
+
"loss": 4.9394378662109375,
|
| 53 |
+
"step": 100
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 0.05390835579514825,
|
| 57 |
+
"eval_loss": 4.784651279449463,
|
| 58 |
+
"eval_runtime": 8.8621,
|
| 59 |
+
"eval_samples_per_second": 1075.032,
|
| 60 |
+
"eval_steps_per_second": 2.144,
|
| 61 |
+
"step": 100
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"epoch": 0.0646900269541779,
|
| 65 |
+
"grad_norm": 0.58203125,
|
| 66 |
+
"learning_rate": 0.001,
|
| 67 |
+
"loss": 4.660997772216797,
|
| 68 |
+
"step": 120
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.07547169811320754,
|
| 72 |
+
"grad_norm": 1.5546875,
|
| 73 |
+
"learning_rate": 0.001,
|
| 74 |
+
"loss": 4.43031005859375,
|
| 75 |
+
"step": 140
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.08086253369272237,
|
| 79 |
+
"eval_loss": 4.252280235290527,
|
| 80 |
+
"eval_runtime": 8.9869,
|
| 81 |
+
"eval_samples_per_second": 1060.1,
|
| 82 |
+
"eval_steps_per_second": 2.114,
|
| 83 |
+
"step": 150
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 0.0862533692722372,
|
| 87 |
+
"grad_norm": 0.48828125,
|
| 88 |
+
"learning_rate": 0.001,
|
| 89 |
+
"loss": 4.2591907501220705,
|
| 90 |
+
"step": 160
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 0.09703504043126684,
|
| 94 |
+
"grad_norm": 0.91015625,
|
| 95 |
+
"learning_rate": 0.001,
|
| 96 |
+
"loss": 4.122463607788086,
|
| 97 |
+
"step": 180
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"epoch": 0.1078167115902965,
|
| 101 |
+
"grad_norm": 0.435546875,
|
| 102 |
+
"learning_rate": 0.001,
|
| 103 |
+
"loss": 3.9880271911621095,
|
| 104 |
+
"step": 200
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.1078167115902965,
|
| 108 |
+
"eval_loss": 3.9186038970947266,
|
| 109 |
+
"eval_runtime": 8.9327,
|
| 110 |
+
"eval_samples_per_second": 1066.53,
|
| 111 |
+
"eval_steps_per_second": 2.127,
|
| 112 |
+
"step": 200
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 0.11859838274932614,
|
| 116 |
+
"grad_norm": 0.73828125,
|
| 117 |
+
"learning_rate": 0.001,
|
| 118 |
+
"loss": 3.8731555938720703,
|
| 119 |
+
"step": 220
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 0.1293800539083558,
|
| 123 |
+
"grad_norm": 0.40234375,
|
| 124 |
+
"learning_rate": 0.001,
|
| 125 |
+
"loss": 3.786928176879883,
|
| 126 |
+
"step": 240
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 0.1347708894878706,
|
| 130 |
+
"eval_loss": 3.682241916656494,
|
| 131 |
+
"eval_runtime": 8.9206,
|
| 132 |
+
"eval_samples_per_second": 1067.975,
|
| 133 |
+
"eval_steps_per_second": 2.13,
|
| 134 |
+
"step": 250
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.14016172506738545,
|
| 138 |
+
"grad_norm": 0.84375,
|
| 139 |
+
"learning_rate": 0.001,
|
| 140 |
+
"loss": 3.6825389862060547,
|
| 141 |
+
"step": 260
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.1509433962264151,
|
| 145 |
+
"grad_norm": 0.51171875,
|
| 146 |
+
"learning_rate": 0.001,
|
| 147 |
+
"loss": 3.6361549377441404,
|
| 148 |
+
"step": 280
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 0.16172506738544473,
|
| 152 |
+
"grad_norm": 1.5703125,
|
| 153 |
+
"learning_rate": 0.001,
|
| 154 |
+
"loss": 3.551879119873047,
|
| 155 |
+
"step": 300
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 0.16172506738544473,
|
| 159 |
+
"eval_loss": 3.5237605571746826,
|
| 160 |
+
"eval_runtime": 9.0532,
|
| 161 |
+
"eval_samples_per_second": 1052.333,
|
| 162 |
+
"eval_steps_per_second": 2.099,
|
| 163 |
+
"step": 300
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.1725067385444744,
|
| 167 |
+
"grad_norm": 0.62109375,
|
| 168 |
+
"learning_rate": 0.001,
|
| 169 |
+
"loss": 3.526197814941406,
|
| 170 |
+
"step": 320
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.18328840970350405,
|
| 174 |
+
"grad_norm": 0.34765625,
|
| 175 |
+
"learning_rate": 0.001,
|
| 176 |
+
"loss": 3.44549446105957,
|
| 177 |
+
"step": 340
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.18867924528301888,
|
| 181 |
+
"eval_loss": 3.4025301933288574,
|
| 182 |
+
"eval_runtime": 9.1498,
|
| 183 |
+
"eval_samples_per_second": 1041.23,
|
| 184 |
+
"eval_steps_per_second": 2.077,
|
| 185 |
+
"step": 350
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.1940700808625337,
|
| 189 |
+
"grad_norm": 1.0234375,
|
| 190 |
+
"learning_rate": 0.001,
|
| 191 |
+
"loss": 3.4011425018310546,
|
| 192 |
+
"step": 360
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.20485175202156333,
|
| 196 |
+
"grad_norm": 0.6171875,
|
| 197 |
+
"learning_rate": 0.001,
|
| 198 |
+
"loss": 3.3427574157714846,
|
| 199 |
+
"step": 380
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.215633423180593,
|
| 203 |
+
"grad_norm": 0.5859375,
|
| 204 |
+
"learning_rate": 0.001,
|
| 205 |
+
"loss": 3.2958106994628906,
|
| 206 |
+
"step": 400
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.215633423180593,
|
| 210 |
+
"eval_loss": 3.2790729999542236,
|
| 211 |
+
"eval_runtime": 9.1929,
|
| 212 |
+
"eval_samples_per_second": 1036.342,
|
| 213 |
+
"eval_steps_per_second": 2.067,
|
| 214 |
+
"step": 400
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"epoch": 0.22641509433962265,
|
| 218 |
+
"grad_norm": 0.5234375,
|
| 219 |
+
"learning_rate": 0.001,
|
| 220 |
+
"loss": 3.2616317749023436,
|
| 221 |
+
"step": 420
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.2371967654986523,
|
| 225 |
+
"grad_norm": 0.5703125,
|
| 226 |
+
"learning_rate": 0.001,
|
| 227 |
+
"loss": 3.2149822235107424,
|
| 228 |
+
"step": 440
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 0.24258760107816713,
|
| 232 |
+
"eval_loss": 3.181051015853882,
|
| 233 |
+
"eval_runtime": 9.1212,
|
| 234 |
+
"eval_samples_per_second": 1044.485,
|
| 235 |
+
"eval_steps_per_second": 2.083,
|
| 236 |
+
"step": 450
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.24797843665768193,
|
| 240 |
+
"grad_norm": 1.0546875,
|
| 241 |
+
"learning_rate": 0.001,
|
| 242 |
+
"loss": 3.1823314666748046,
|
| 243 |
+
"step": 460
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.2587601078167116,
|
| 247 |
+
"grad_norm": 0.5390625,
|
| 248 |
+
"learning_rate": 0.001,
|
| 249 |
+
"loss": 3.1341114044189453,
|
| 250 |
+
"step": 480
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 0.2695417789757412,
|
| 254 |
+
"grad_norm": 0.9375,
|
| 255 |
+
"learning_rate": 0.001,
|
| 256 |
+
"loss": 3.101780128479004,
|
| 257 |
+
"step": 500
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 0.2695417789757412,
|
| 261 |
+
"eval_loss": 3.1055757999420166,
|
| 262 |
+
"eval_runtime": 9.2225,
|
| 263 |
+
"eval_samples_per_second": 1033.013,
|
| 264 |
+
"eval_steps_per_second": 2.06,
|
| 265 |
+
"step": 500
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.2803234501347709,
|
| 269 |
+
"grad_norm": 0.5078125,
|
| 270 |
+
"learning_rate": 0.001,
|
| 271 |
+
"loss": 3.0730880737304687,
|
| 272 |
+
"step": 520
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.29110512129380056,
|
| 276 |
+
"grad_norm": 0.609375,
|
| 277 |
+
"learning_rate": 0.001,
|
| 278 |
+
"loss": 3.0466753005981446,
|
| 279 |
+
"step": 540
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.29649595687331537,
|
| 283 |
+
"eval_loss": 3.022505283355713,
|
| 284 |
+
"eval_runtime": 9.1838,
|
| 285 |
+
"eval_samples_per_second": 1037.366,
|
| 286 |
+
"eval_steps_per_second": 2.069,
|
| 287 |
+
"step": 550
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 0.3018867924528302,
|
| 291 |
+
"grad_norm": 0.5,
|
| 292 |
+
"learning_rate": 0.001,
|
| 293 |
+
"loss": 3.0200361251831054,
|
| 294 |
+
"step": 560
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 0.31266846361185985,
|
| 298 |
+
"grad_norm": 0.73046875,
|
| 299 |
+
"learning_rate": 0.001,
|
| 300 |
+
"loss": 2.9901798248291014,
|
| 301 |
+
"step": 580
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 0.32345013477088946,
|
| 305 |
+
"grad_norm": 0.73046875,
|
| 306 |
+
"learning_rate": 0.001,
|
| 307 |
+
"loss": 2.958896255493164,
|
| 308 |
+
"step": 600
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.32345013477088946,
|
| 312 |
+
"eval_loss": 2.95377516746521,
|
| 313 |
+
"eval_runtime": 9.1618,
|
| 314 |
+
"eval_samples_per_second": 1039.857,
|
| 315 |
+
"eval_steps_per_second": 2.074,
|
| 316 |
+
"step": 600
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 0.33423180592991913,
|
| 320 |
+
"grad_norm": 0.78125,
|
| 321 |
+
"learning_rate": 0.001,
|
| 322 |
+
"loss": 2.952092933654785,
|
| 323 |
+
"step": 620
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 0.3450134770889488,
|
| 327 |
+
"grad_norm": 0.44921875,
|
| 328 |
+
"learning_rate": 0.001,
|
| 329 |
+
"loss": 2.9132663726806642,
|
| 330 |
+
"step": 640
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 0.3504043126684636,
|
| 334 |
+
"eval_loss": 2.905498504638672,
|
| 335 |
+
"eval_runtime": 9.1525,
|
| 336 |
+
"eval_samples_per_second": 1040.913,
|
| 337 |
+
"eval_steps_per_second": 2.076,
|
| 338 |
+
"step": 650
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.3557951482479784,
|
| 342 |
+
"grad_norm": 0.98828125,
|
| 343 |
+
"learning_rate": 0.001,
|
| 344 |
+
"loss": 2.900470161437988,
|
| 345 |
+
"step": 660
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.3665768194070081,
|
| 349 |
+
"grad_norm": 0.56640625,
|
| 350 |
+
"learning_rate": 0.001,
|
| 351 |
+
"loss": 2.8784067153930666,
|
| 352 |
+
"step": 680
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.37735849056603776,
|
| 356 |
+
"grad_norm": 0.92578125,
|
| 357 |
+
"learning_rate": 0.001,
|
| 358 |
+
"loss": 2.8517574310302733,
|
| 359 |
+
"step": 700
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.37735849056603776,
|
| 363 |
+
"eval_loss": 2.868436336517334,
|
| 364 |
+
"eval_runtime": 9.2031,
|
| 365 |
+
"eval_samples_per_second": 1035.193,
|
| 366 |
+
"eval_steps_per_second": 2.065,
|
| 367 |
+
"step": 700
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.3881401617250674,
|
| 371 |
+
"grad_norm": 0.486328125,
|
| 372 |
+
"learning_rate": 0.001,
|
| 373 |
+
"loss": 2.846698760986328,
|
| 374 |
+
"step": 720
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.39892183288409705,
|
| 378 |
+
"grad_norm": 0.82421875,
|
| 379 |
+
"learning_rate": 0.001,
|
| 380 |
+
"loss": 2.8154937744140627,
|
| 381 |
+
"step": 740
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.40431266846361186,
|
| 385 |
+
"eval_loss": 2.813324451446533,
|
| 386 |
+
"eval_runtime": 9.1186,
|
| 387 |
+
"eval_samples_per_second": 1044.786,
|
| 388 |
+
"eval_steps_per_second": 2.084,
|
| 389 |
+
"step": 750
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 0.40970350404312667,
|
| 393 |
+
"grad_norm": 0.54296875,
|
| 394 |
+
"learning_rate": 0.001,
|
| 395 |
+
"loss": 2.8071964263916014,
|
| 396 |
+
"step": 760
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 0.42048517520215634,
|
| 400 |
+
"grad_norm": 0.69140625,
|
| 401 |
+
"learning_rate": 0.001,
|
| 402 |
+
"loss": 2.7811683654785155,
|
| 403 |
+
"step": 780
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 0.431266846361186,
|
| 407 |
+
"grad_norm": 0.73828125,
|
| 408 |
+
"learning_rate": 0.001,
|
| 409 |
+
"loss": 2.782184028625488,
|
| 410 |
+
"step": 800
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.431266846361186,
|
| 414 |
+
"eval_loss": 2.778449535369873,
|
| 415 |
+
"eval_runtime": 9.2351,
|
| 416 |
+
"eval_samples_per_second": 1031.606,
|
| 417 |
+
"eval_steps_per_second": 2.057,
|
| 418 |
+
"step": 800
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 0.4420485175202156,
|
| 422 |
+
"grad_norm": 0.44921875,
|
| 423 |
+
"learning_rate": 0.001,
|
| 424 |
+
"loss": 2.754811668395996,
|
| 425 |
+
"step": 820
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 0.4528301886792453,
|
| 429 |
+
"grad_norm": 0.89453125,
|
| 430 |
+
"learning_rate": 0.001,
|
| 431 |
+
"loss": 2.7389667510986326,
|
| 432 |
+
"step": 840
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 0.4582210242587601,
|
| 436 |
+
"eval_loss": 2.736283540725708,
|
| 437 |
+
"eval_runtime": 9.2108,
|
| 438 |
+
"eval_samples_per_second": 1034.331,
|
| 439 |
+
"eval_steps_per_second": 2.063,
|
| 440 |
+
"step": 850
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.4636118598382749,
|
| 444 |
+
"grad_norm": 0.53515625,
|
| 445 |
+
"learning_rate": 0.001,
|
| 446 |
+
"loss": 2.728610801696777,
|
| 447 |
+
"step": 860
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.4743935309973046,
|
| 451 |
+
"grad_norm": 0.98828125,
|
| 452 |
+
"learning_rate": 0.001,
|
| 453 |
+
"loss": 2.703776550292969,
|
| 454 |
+
"step": 880
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 0.48517520215633425,
|
| 458 |
+
"grad_norm": 0.75390625,
|
| 459 |
+
"learning_rate": 0.001,
|
| 460 |
+
"loss": 2.712217330932617,
|
| 461 |
+
"step": 900
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 0.48517520215633425,
|
| 465 |
+
"eval_loss": 2.7041544914245605,
|
| 466 |
+
"eval_runtime": 9.4718,
|
| 467 |
+
"eval_samples_per_second": 1005.83,
|
| 468 |
+
"eval_steps_per_second": 2.006,
|
| 469 |
+
"step": 900
|
| 470 |
+
}
|
| 471 |
+
],
|
| 472 |
+
"logging_steps": 20,
|
| 473 |
+
"max_steps": 1000,
|
| 474 |
+
"num_input_tokens_seen": 0,
|
| 475 |
+
"num_train_epochs": 1,
|
| 476 |
+
"save_steps": 100,
|
| 477 |
+
"stateful_callbacks": {
|
| 478 |
+
"TrainerControl": {
|
| 479 |
+
"args": {
|
| 480 |
+
"should_epoch_stop": false,
|
| 481 |
+
"should_evaluate": false,
|
| 482 |
+
"should_log": false,
|
| 483 |
+
"should_save": true,
|
| 484 |
+
"should_training_stop": false
|
| 485 |
+
},
|
| 486 |
+
"attributes": {}
|
| 487 |
+
}
|
| 488 |
+
},
|
| 489 |
+
"total_flos": 4878284252774400.0,
|
| 490 |
+
"train_batch_size": 512,
|
| 491 |
+
"trial_name": null,
|
| 492 |
+
"trial_params": null
|
| 493 |
+
}
|
out/mlp-linear-21L_run/checkpoint-900/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f710638f98cb9aa723847c3078af1df019ed60911a4c32cfe8698e762841d517
|
| 3 |
+
size 4920
|
out/mlp-linear-21L_run/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 7959296
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:23ae85e4a038e289ca876dca6ccf4fd8a2a5cc4c9d17dd3a1d0dfe677c6fb506
|
| 3 |
size 7959296
|
out/mlp-linear-21L_run/training_log.jsonl
CHANGED
|
@@ -41,3 +41,27 @@
|
|
| 41 |
{"step": 600, "epoch": 0.32345013477088946, "timestamp": 1786140343.349682, "loss": 2.958896255493164, "grad_norm": 0.73046875, "learning_rate": 0.001, "train/total_time_seconds": 190.6581477150321, "train/time_per_step_avg": 0.31607903823256495, "train/epoch_time_elapsed": 521.0774472914636, "train/estimated_remaining_minutes": 2.1184238635003565}
|
| 42 |
{"step": 600, "epoch": 0.32345013477088946, "timestamp": 1786140352.5130959, "eval_loss": 2.95377516746521, "eval_runtime": 9.1618, "eval_samples_per_second": 1039.857, "eval_steps_per_second": 2.074, "train/total_time_seconds": 190.6581477150321, "train/time_per_step_avg": 0.31607903823256495, "train/epoch_time_elapsed": 530.2408594228327, "train/estimated_remaining_minutes": 2.1184238635003565}
|
| 43 |
{"step": 620, "epoch": 0.33423180592991913, "timestamp": 1786140366.5955102, "loss": 2.952092933654785, "grad_norm": 0.78125, "learning_rate": 0.001, "train/total_time_seconds": 196.97960712760687, "train/time_per_step_avg": 0.31603967782109976, "train/epoch_time_elapsed": 544.323274012655, "train/estimated_remaining_minutes": 2.0121572771099623}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 41 |
{"step": 600, "epoch": 0.32345013477088946, "timestamp": 1786140343.349682, "loss": 2.958896255493164, "grad_norm": 0.73046875, "learning_rate": 0.001, "train/total_time_seconds": 190.6581477150321, "train/time_per_step_avg": 0.31607903823256495, "train/epoch_time_elapsed": 521.0774472914636, "train/estimated_remaining_minutes": 2.1184238635003565}
|
| 42 |
{"step": 600, "epoch": 0.32345013477088946, "timestamp": 1786140352.5130959, "eval_loss": 2.95377516746521, "eval_runtime": 9.1618, "eval_samples_per_second": 1039.857, "eval_steps_per_second": 2.074, "train/total_time_seconds": 190.6581477150321, "train/time_per_step_avg": 0.31607903823256495, "train/epoch_time_elapsed": 530.2408594228327, "train/estimated_remaining_minutes": 2.1184238635003565}
|
| 43 |
{"step": 620, "epoch": 0.33423180592991913, "timestamp": 1786140366.5955102, "loss": 2.952092933654785, "grad_norm": 0.78125, "learning_rate": 0.001, "train/total_time_seconds": 196.97960712760687, "train/time_per_step_avg": 0.31603967782109976, "train/epoch_time_elapsed": 544.323274012655, "train/estimated_remaining_minutes": 2.0121572771099623}
|
| 44 |
+
{"step": 640, "epoch": 0.3450134770889488, "timestamp": 1786140380.663665, "loss": 2.9132663726806642, "grad_norm": 0.44921875, "learning_rate": 0.001, "train/total_time_seconds": 203.29968735203147, "train/time_per_step_avg": 0.3160211337730289, "train/epoch_time_elapsed": 558.3914296738803, "train/estimated_remaining_minutes": 1.905934568925295}
|
| 45 |
+
{"step": 650, "epoch": 0.3504043126684636, "timestamp": 1786140396.8474813, "eval_loss": 2.905498504638672, "eval_runtime": 9.1525, "eval_samples_per_second": 1040.913, "eval_steps_per_second": 2.076, "train/total_time_seconds": 206.45854242891073, "train/time_per_step_avg": 0.3159821989759803, "train/epoch_time_elapsed": 574.5752436704934, "train/estimated_remaining_minutes": 1.8528330730799683}
|
| 46 |
+
{"step": 660, "epoch": 0.3557951482479784, "timestamp": 1786140403.8380544, "loss": 2.900470161437988, "grad_norm": 0.98828125, "learning_rate": 0.001, "train/total_time_seconds": 209.62075051665306, "train/time_per_step_avg": 0.3160286565870047, "train/epoch_time_elapsed": 581.5658187195659, "train/estimated_remaining_minutes": 1.7997741205975262}
|
| 47 |
+
{"step": 680, "epoch": 0.3665768194070081, "timestamp": 1786140417.9181857, "loss": 2.8784067153930666, "grad_norm": 0.56640625, "learning_rate": 0.001, "train/total_time_seconds": 215.93968116119504, "train/time_per_step_avg": 0.31599062215536833, "train/epoch_time_elapsed": 595.6459505409002, "train/estimated_remaining_minutes": 1.6936445581270196}
|
| 48 |
+
{"step": 700, "epoch": 0.37735849056603776, "timestamp": 1786140431.9665506, "loss": 2.8517574310302733, "grad_norm": 0.92578125, "learning_rate": 0.001, "train/total_time_seconds": 222.25654493644834, "train/time_per_step_avg": 0.31598397221416236, "train/epoch_time_elapsed": 609.6943158768117, "train/estimated_remaining_minutes": 1.5875467495460596}
|
| 49 |
+
{"step": 700, "epoch": 0.37735849056603776, "timestamp": 1786140441.17121, "eval_loss": 2.868436336517334, "eval_runtime": 9.2031, "eval_samples_per_second": 1035.193, "eval_steps_per_second": 2.065, "train/total_time_seconds": 222.25654493644834, "train/time_per_step_avg": 0.31598397221416236, "train/epoch_time_elapsed": 618.898972183466, "train/estimated_remaining_minutes": 1.5875467495460596}
|
| 50 |
+
{"step": 720, "epoch": 0.3881401617250674, "timestamp": 1786140455.3046455, "loss": 2.846698760986328, "grad_norm": 0.486328125, "learning_rate": 0.001, "train/total_time_seconds": 228.57606745883822, "train/time_per_step_avg": 0.3159646033123136, "train/epoch_time_elapsed": 633.0324101299047, "train/estimated_remaining_minutes": 1.481511548344322}
|
| 51 |
+
{"step": 740, "epoch": 0.39892183288409705, "timestamp": 1786140469.2788513, "loss": 2.8154937744140627, "grad_norm": 0.82421875, "learning_rate": 0.001, "train/total_time_seconds": 234.8880673572421, "train/time_per_step_avg": 0.31588380005210637, "train/epoch_time_elapsed": 647.006616409868, "train/estimated_remaining_minutes": 1.375470664704571}
|
| 52 |
+
{"step": 750, "epoch": 0.40431266846361186, "timestamp": 1786140485.4477496, "eval_loss": 2.813324451446533, "eval_runtime": 9.1186, "eval_samples_per_second": 1044.786, "eval_steps_per_second": 2.084, "train/total_time_seconds": 238.04644763469696, "train/time_per_step_avg": 0.31587905205786226, "train/epoch_time_elapsed": 663.1755127944052, "train/estimated_remaining_minutes": 1.3224802646372056}
|
| 53 |
+
{"step": 760, "epoch": 0.40970350404312667, "timestamp": 1786140492.533495, "loss": 2.8071964263916014, "grad_norm": 0.54296875, "learning_rate": 0.001, "train/total_time_seconds": 241.20690286532044, "train/time_per_step_avg": 0.31586152348667385, "train/epoch_time_elapsed": 670.2612607181072, "train/estimated_remaining_minutes": 1.269510015080634}
|
| 54 |
+
{"step": 780, "epoch": 0.42048517520215634, "timestamp": 1786140506.7619843, "loss": 2.7811683654785155, "grad_norm": 0.69140625, "learning_rate": 0.001, "train/total_time_seconds": 247.52511202916503, "train/time_per_step_avg": 0.3158543086796999, "train/epoch_time_elapsed": 684.489748865366, "train/estimated_remaining_minutes": 1.1635795864618868}
|
| 55 |
+
{"step": 800, "epoch": 0.431266846361186, "timestamp": 1786140520.9534838, "loss": 2.782184028625488, "grad_norm": 0.73828125, "learning_rate": 0.001, "train/total_time_seconds": 253.83729097992182, "train/time_per_step_avg": 0.31580746043473484, "train/epoch_time_elapsed": 698.6812483593822, "train/estimated_remaining_minutes": 1.0576553790830077}
|
| 56 |
+
{"step": 800, "epoch": 0.431266846361186, "timestamp": 1786140530.190191, "eval_loss": 2.778449535369873, "eval_runtime": 9.2351, "eval_samples_per_second": 1031.606, "eval_steps_per_second": 2.057, "train/total_time_seconds": 253.83729097992182, "train/time_per_step_avg": 0.31580746043473484, "train/epoch_time_elapsed": 707.9179544597864, "train/estimated_remaining_minutes": 1.0576553790830077}
|
| 57 |
+
{"step": 820, "epoch": 0.4420485175202156, "timestamp": 1786140544.3385065, "loss": 2.754811668395996, "grad_norm": 0.44921875, "learning_rate": 0.001, "train/total_time_seconds": 260.1481362618506, "train/time_per_step_avg": 0.3157206880301237, "train/epoch_time_elapsed": 722.0662711746991, "train/estimated_remaining_minutes": 0.9517614741287217}
|
| 58 |
+
{"step": 840, "epoch": 0.4528301886792453, "timestamp": 1786140558.3534312, "loss": 2.7389667510986326, "grad_norm": 0.89453125, "learning_rate": 0.001, "train/total_time_seconds": 266.4599728770554, "train/time_per_step_avg": 0.315719055198133, "train/epoch_time_elapsed": 736.0811963900924, "train/estimated_remaining_minutes": 0.8459046758001759}
|
| 59 |
+
{"step": 850, "epoch": 0.4582210242587601, "timestamp": 1786140574.6184158, "eval_loss": 2.736283540725708, "eval_runtime": 9.2108, "eval_samples_per_second": 1034.331, "eval_steps_per_second": 2.063, "train/total_time_seconds": 269.6180298104882, "train/time_per_step_avg": 0.31571582175791263, "train/epoch_time_elapsed": 752.3461726717651, "train/estimated_remaining_minutes": 0.7929942053249655}
|
| 60 |
+
{"step": 860, "epoch": 0.4636118598382749, "timestamp": 1786140581.8180475, "loss": 2.728610801696777, "grad_norm": 0.53515625, "learning_rate": 0.001, "train/total_time_seconds": 272.7832880653441, "train/time_per_step_avg": 0.3157638520002365, "train/epoch_time_elapsed": 759.5458110570908, "train/estimated_remaining_minutes": 0.7401096963013212}
|
| 61 |
+
{"step": 880, "epoch": 0.4743935309973046, "timestamp": 1786140596.0643246, "loss": 2.703776550292969, "grad_norm": 0.98828125, "learning_rate": 0.001, "train/total_time_seconds": 279.11409997940063, "train/time_per_step_avg": 0.31588987950235603, "train/epoch_time_elapsed": 773.7920881733298, "train/estimated_remaining_minutes": 0.6343502272259105}
|
| 62 |
+
{"step": 900, "epoch": 0.48517520215633425, "timestamp": 1786140610.5076487, "loss": 2.712217330932617, "grad_norm": 0.75390625, "learning_rate": 0.001, "train/total_time_seconds": 285.4424224048853, "train/time_per_step_avg": 0.31605131424963473, "train/epoch_time_elapsed": 788.2354110814631, "train/estimated_remaining_minutes": 0.5285970785275654}
|
| 63 |
+
{"step": 900, "epoch": 0.48517520215633425, "timestamp": 1786140619.9812293, "eval_loss": 2.7041544914245605, "eval_runtime": 9.4718, "eval_samples_per_second": 1005.83, "eval_steps_per_second": 2.006, "train/total_time_seconds": 285.4424224048853, "train/time_per_step_avg": 0.31605131424963473, "train/epoch_time_elapsed": 797.7089912705123, "train/estimated_remaining_minutes": 0.5285970785275654}
|
| 64 |
+
{"step": 920, "epoch": 0.49595687331536387, "timestamp": 1786140634.3566592, "loss": 2.677124786376953, "grad_norm": 0.48828125, "learning_rate": 0.001, "train/total_time_seconds": 291.7641347832978, "train/time_per_step_avg": 0.3161599852144718, "train/epoch_time_elapsed": 812.0844232290983, "train/estimated_remaining_minutes": 0.42284657214970695}
|
| 65 |
+
{"step": 940, "epoch": 0.5067385444743935, "timestamp": 1786140648.6530638, "loss": 2.6723052978515627, "grad_norm": 0.9921875, "learning_rate": 0.001, "train/total_time_seconds": 298.08655486628413, "train/time_per_step_avg": 0.31626581989228725, "train/epoch_time_elapsed": 826.3808276876807, "train/estimated_remaining_minutes": 0.3171133562407278}
|
| 66 |
+
{"step": 950, "epoch": 0.5121293800539084, "timestamp": 1786140665.325624, "eval_loss": 2.6709580421447754, "eval_runtime": 9.5201, "eval_samples_per_second": 1000.726, "eval_steps_per_second": 1.996, "train/total_time_seconds": 301.25240771844983, "train/time_per_step_avg": 0.3163437790796161, "train/epoch_time_elapsed": 843.0533857010305, "train/estimated_remaining_minutes": 0.2642564979986402}
|
| 67 |
+
{"step": 960, "epoch": 0.5175202156334232, "timestamp": 1786140672.499597, "loss": 2.672710418701172, "grad_norm": 0.482421875, "learning_rate": 0.001, "train/total_time_seconds": 304.4166898280382, "train/time_per_step_avg": 0.3163340176269412, "train/epoch_time_elapsed": 850.2273597568274, "train/estimated_remaining_minutes": 0.21140047904724876}
|